Hopefully it won't break again. It seems there is a database field called "error_log" which just... keeps filling up with random errors until the database gets filled up and completely crashes. I *think* I fixed it, or at least slowed it down. At the very least, I'll keep an eye out on that database.
